In this work, we empirically evaluated expert-led debunking of health misinformation on TikTok with n=420 survey and n=20 interview participants. Unlike fact-checkers, health professionals debunk misinformation non-anonymously through video-against-video formats (stitching/”duetting”), rather than using labels. We analyzed 5,161 such posts to select six misinformation and six debunking videos across three common topics – two general health, two mental health, and two nutrition – for statistical comparison. Participants exposed to debunking videos believed misinformation claims significantly less than those exposed to misinformation videos in all six conditions. Experts were seen as more credible than misinformation creators, except in one instance involving mental health. Thematic analysis showed that expert-led debunking succeeded because experts’ videos aligned with the Debunking Handbook method for effective refutation. Experts’ credibility is derived mainly from being perceived as non-typical influencers who maintain reputable TikTok personas by providing qualified medical evidence and advice.

increasingly realistic, making it possible to replicate a person's voice with just a few seconds of audio. Cloned voices are already being used in impersonation scams, including family emergency schemes where callers use cloned voices to pose as relatives in distress to solicit money. Postmortem voice cloning introduces new opportunities for misuse, allowing bad actors to impersonate deceased individuals or emotionally manipulate surviving family members. My dissertation research examines the risks that individuals associate with voice cloning. Using a mixed-methods approach that incorporates surveys and participant interviews, I will also explore assumptions about voice privacy and expectations regarding regulation and prevention of unauthorized use. I will use thematic analysis that is guided by an established voice cloning harm taxonomy. The goal of my research is to understand the perspectives of individuals to inform safeguards, legal protections, and platform responsibilities surrounding the unauthorized use of a person's voice after their death.

Telephone scams often attempt to defraud or steal the identity of individuals by eliciting a response to a pressing request for payments or submission of information such as social security numbers. These calls usually come from numbers that individuals have not encountered before, and the request is passed through as a pre-recorded message with a generic voice. But AI-enabled voice cloning and the ability to spoof Caller IDs have given scammers the opportunity to run schemes targeted individuals with a cloned familiar voice coming from trusted numbers. This paper reports the findings from an empirical study that replicates this scenario with 14 participants (7 pairs of family relatives of friends) to capture the experiences of receiving such a scam call and responses to it. The results of our thematic analysis show that the familiar voice, coming from a trusted number of a family relative or a friend, is highly persuasive towards deceiving the call receiver (i.e., callee) to indeed follow through with the scammer’s request. The callee, together with the caller or the participants who volunteered their voices for cloning, saw this scam working particularly in the context of family emergencies, as real-world reports have surfaced about children held for ransom, grandchildren under arrest, or relatives in car accidents. The callers and callees saw no immediate way to fend off these “family emergency scams” than for families and friends to work on “family/friend codewords” that are hard to be inferred by the scammers (and thus cloned). We discuss our findings towards the development of user-centered interventions that would facilitate the detection of a wide range of AI-enabled voice cloning scams, in addition to the suggested personal ways of scam detection.

This paper reports on a study exploring how two groups of individuals, legally blind$(n=36)$and sighted ones$(n=36)$, react to aural telephone scam warnings in naturalistic settings. As spoofing a CallerID is trivial, communicating the context of an incoming call instead offers a better possibility to warn a receiver about a potential scam. Usually, such warnings are visual in nature and fail to cater to users with visual disabilities. To address this exclusion, we developed an aural variant of telephone scam warnings and tested them in three conditions: baseline (no warning), short warning, and contextual warning that preceded the scam's content. We tested the two most common scam scenarios: fraud (interest rate reduction) and identity theft (social security number) by coldcalling participants and recording their actions, and debriefing and obtaining consent afterward. Only two participants “pressed one” as the scam demanded, both from the legally blind group that heard the contextual warning for the social security scenario. Upon close inspection, we learned that one of them did so because of accessibility issues with their screen reader and the other did so intentionally because the warning convinced them to waste the scammer's time, so they don't scam vulnerable people. Both the legally blind and sighted participants found the contextual warnings as powerful usable security cues that, together with STIR/SHAKEN indicators like Scam Likely, would provide robust protection against any type of scam. We also discussed the potential privacy implications of the contextual warnings and collected recommendations for usably accessible implementation.

Older adults habitually encounter misinformation, yet little is known about their experiences with it. In this study, we employed a mixed-methods approach, combining a survey (n=119) with semi-structured interviews ( n =21), to investigate how older adults in America conceptualize, discern, and contextualize social media misinformation. Given the historical context of misinformation being used to influence voting outcomes, our study specifically examined this phenomenon from a voting intention perspective. Our findings reveal that 62% of participants intending to vote Democrat perceived a manipulative political purpose behind the spread of misinformation, whereas only 5% of those intending to vote Republican believed that misinformation serves a political dissent purpose. Regardless of voting intentions, most participants relied on source heuristics and fact-checking to discern truth from misinformation on social media. A major concern among participants was the biased reasoning influenced by personal values and emotions affected by misinformation. Notably, 74% of participants intending to vote Democrat were concerned that misinformation would escalate extremism in the future. In contrast, those intending to vote Republican, those undecided, or those planning to abstain expressed concerns that misinformation would further erode trust in democratic institutions, particularly in public health and free and fair elections. During our interviews, we discovered that 63% of participants intending to vote Republican mentioned that Republican or conservative voices often disseminate misinformation, even though these participants were closely aligned with this political ideology.

Audio deepfakes are a form of deception where convincing speech sentences are synthesized through machine learning means to give an impression of a human speaker. Audio deepfakes emerge as an attractive vector for targeting users that rely on audio accessibility, such as individuals who are blind or low vision. The critical reliance on speech both as a medium and an affordance puts this population at an undue risk of being deceived as they rely solely on themselves to detect whether a piece of audio is a deepfake or not. To better understand the nature of this risk considering the nuanced reliance on assistive technologies such as screen readers, we conducted a user study with n=16 blind and low vision individuals from the US. Our participants achieved an overall discernment accuracy of 59%, and clips identified as deep fakes were only actually deepfakes in 50.8% of the cases (precision). The participants that self-identified as "low vision" performed slightly better (accuracy of 61%, precision of 64%) compared to the ones that self-identified as "blind" (accuracy of 55%, precision of 56%). Our qualitative results show that the participants in the "blind" group mostly considered a combination of infliction, imperfections in the voice, and the intensity in the speech delivery as discernment factors. The participants in the "low vision" group mostly used the speaker's pitch, enunciation, emotion, and the fluency and articulation of the speaker as discernment cues. Overall, participants felt that audio deepfakes have the potential to deceive visually impaired individuals with political disinformation, impersonate their voice in authentication and smart homes, and specifically target them with voice phishing and enhanced scams.

Children encounter misinformation on social media in a similar capacity as their parents. Unlike their parents, children are an exceptionally vulnerable population because their cognitive abilities and emotional regulation are still maturing, rendering them more susceptible to misinformation and falsehoods online. Yet, little is known about children’s experience with misinformation as well as what their parents think of the misinformation’s effect on child development. To answer these questions, we combined a qualitative survey of parents (n=87) with semi-structured interviews of both parents and children (n=12). We found that children identify misinformation as content used to trick people on social media, such as deep fakes, memes with political context, or celebrity/influencer rumors. Children revealed they ask Siri whether a social media video or post is created to trick them before they search on Google or ask their parents about its accuracy. Parents expressed discontent that their children are impressionable to misinformation, stating that the burden falls on them to help their children develop critical thinking skills for navigating falsehoods on social media. Here, the majority of parents felt that schools should also teach these skills as well as media literacy to their children. Misinformation, according to both parents and children, affects the family relationships especially with grandparents with different political views than theirs.